Agents, Infrastructure,
Applications and Norms
Michael Luck
University of Southampton, UK
 Monday
• Agents for next generation computing
 AgentLink Roadmap
 Tuesday
• The case for agents
• Agent Infrastructure
 Conceptual: SMART
 Technical: Paradigma/actSMART
• Agents and Bioinformatics
 GeneWeaver
 myGrid
 Wednesday
• Norms
• Pitfalls
Agent Technology: Enabling
Next Generation Computing
A Roadmap for Agent Based Computing
Michael Luck, University of Southampton, UK
What are agents?
AgentLink and the Roadmap
Current state-of-the-art
Short, medium and long-term predictions
Technical challenges
Community challenges
Application Opportunities
What is an agent?
 A computer system capable of flexible,
autonomous (problem-solving) action,
situated in dynamic, open,
unpredictable and typically multi-agent
What is an agent?
 A computer system capable of flexible,
autonomous (problem-solving) action,
situated in dynamic, open,
unpredictable and typically multi-agent
 control over internal state and over own
What is an agent?
 A computer system capable of flexible,
autonomous (problem-solving) action,
situated in dynamic, open,
unpredictable and typically multi-agent
 experiences environment through
sensors and acts through effectors
What is an agent?
 A computer system capable of flexible,
autonomous (problem-solving) action,
situated in dynamic, open, unpredictable and
typically multi-agent domains.
 reactive: respond in timely fashion to
environmental change
 proactive: act in anticipation of future goals
Multiple Agents
In most cases, single agent is insufficient
• no such thing as a single agent system (!?)
• multiple agents are the norm, to represent:
 natural decentralisation
 multiple loci of control
 multiple perspectives
 competing interests
Agent Interactions
 Interaction between agents is inevitable
• to achieve individual objectives, to manage interdependencies
 Conceptualised as taking place at knowledgelevel
• which goals, at what time, by whom, what for
 Flexible run-time initiation and response
• cf. design-time, hard-wired nature of extant
AgentLink and the Roadmap
What is AgentLink?
 Open network for agent-based
 AgentLink II started in August 2000.
 Intended to give European industry a
head start in a crucial new area of IT.
 Builds on existing activities from
AgentLink (1998-2000)
AgentLink Goals
 Competitive advantage through
promotion of agent systems technology
 Improvement in standard, profile,
industrial relevance of research in
 Promote excellence of teaching and
 High quality forum for R&D
What does AgentLink do?
 Industry action
• gaining advantage for Euro industry
 Research coordination
• excellence & relevance of Euro research
 Education & training
• fostering agent skills
 Special Interest Groups
• focused interactions
 Information infrastructrure
• facilitating AgentLink work
The Roadmap: Aims
 A key deliverable of AgentLink II
 Derives from work of AgentLink SIGs
 Draws on Industry and Research
 Aimed at policy-makers, funding
agencies, academics, industrialists
 Aims to focus future R&D efforts
Special Interest Groups
 Agent-Mediated Electronic Commerce
 Agent-Based Social Simulation
 Methodologies and Software Engineering for
Agent Systems
 Intelligent Information Agents
 Intelligent and Mobile Agents for Telecoms and
the Internet
 Agents that Learn, Adapt and Discover
 Logic and Agents
The Roadmap: Process
 Core roadmapping team:
• Michael Luck
• Peter McBurney
• Chris Preist
Inputs from SIGs: area roadmaps
Specific reviews
Wide consultation exercise
Collation and integration
State of the art
Views of Agents
To support next generation computing
through facilitating agent technologies
 As a metaphor for the design of complex,
distributed computational systems
 As a source of technologies
 As simulation models of complex realworld systems, such as in biology and
Agents as Design
Agent oriented software engineering
Agent architectures
Mobile agents
Agent infrastructure
Electronic institutions
Agent technologies
Multi-agent planning
Agent communication languages
Coordination mechanisms
Matchmaking architectures
Information agents and basic ontologies
Auction mechanism design
Negotiation strategies
Links to other disciplines
Social sciences
Application and Deployment
 Assistant agents
 Multi-agent decision systems
 Multi-agent simulation systems
 IBM, HP Labs, Siemens, Motorola, BT
 Lost Wax, Agent Oriented Software,
Whitestein, Living Systems, iSOCO
The Roadmap Timeline
Sharing of knowledge and goals
Design by same or diverse teams
Languages and interaction protocols
Scale of agents, users, complexity
Design methodologies
Current situation
 One design team
 Agents sharing common goals
 Closed agent systems applied in specific
 Ad-hoc designs
 Predefined communications protocols
and languages
 Scalability only in simulation
Short term to 2005
 Fewer common goals
 Use of semi-structured agent
communication languages (such as FIPA
 Top-down design methodologies such
 Scalability extended to predetermined
and domain-specific environments
Medium term 2006-2008
Design by different teams
Use of agreed protocols and languages
Standard, agent-specific design methodologies
Open agent systems in specific domains (such
as in bioinformatics and e-commerce)
 More general scalability, arbitrary numbers and
diversity of agents in each such domain
 Bridging agents translating between domains
Long Term 2009 Design by diverse teams
 Truly-open and fully-scalable multi-agent
 Across domains
 Agents capable of learning appropriate
communications protocols upon entry to a
 Protocols emerging and evolving through
actual agent interactions.
The Roadmap Timeline
Technological Challenges
Technological Challenges
 Increase quality of agent systems to
industrial standard
 Provide effective agreed standards to
allow open systems development
 Provide infrastructure for open agent
 Develop reasoning capabilities for
agents in open environments
Technological Challenges
 Develop agent ability to adapt to
changes in environment
 Develop agent ability to understand
user requirements
 Ensure user confidence and trust in
Industrial Strength Software
 Fundamental obstacle to take-up is lack of
mature software methodology
• Coordination, interaction, organisation, society joint goals, plans, norms, protocols, etc
• Libraries of …
 agent and organisation models
 communication languages and patterns
 ontology patterns
 CASE tools
 AUML is one example
Industrial Strength Software
Agreed Standards
 FIPA and OMG
• Agent platform architectures
• Semantic communication and content
languages for messages and protocols
• Interoperability
• Ontology modelling
 Public libraries in other areas will be
Agreed Standards
Semantic Infrastructure for Open
 Need to understand relation of agents,
databases and information systems
 Real world implications of information agents
 Benchmarks for performance
 Use new web standards for structural and
semantic description
 Services that make use of such semantic
Semantic Infrastructure for Open
 Ontologies
Timely covergence of technologies
Generic tool and service support
Shared ontologies
Semantic Web community exploring many
Semantic Infrastructure for Open
Reasoning in Open Environments
 Cannot handle issues inherent in open
multi-agent systems
Trust and accountability
Failure handling and recovery
Societal change
 Domain-specific models of reasoning
Reasoning in Open Environments
 Coalition formation
 Dynamic establishment of virtual
 Demanded by emerging computational
infrastructure such as
• Grid
• Web Services
• eBusiness workflow systems
Reasoning in Open Environments
 Negotiation and argumentation
• Some existing work but currently in infancy
 Need to address
Rigorous testing in realistic environments
Overarching theory or methodology
Efficient argumentation engines
Techniques for user preference specification
Techniques for user creation and dissolution of
virtual organisations
Reasoning in Open Environments
Learning Technologies
 Ability to understand user requirements
• Integration of machine learning
• XML profiles
 Ability to adapt to changes in environment
• Multi-agent learning is far behind single agent
• Personal information management raises issues of
 Relationship to Semantic Web
Learning Technologies
Trust and Reputation
 User confidence
 Trust of users in agents
• Issues of autonomy
• Formal methods and verification
 Trust of agents in agents
• Norms
• Reputation
• Contracts
Trust and Reputation
Challenges for the Agent
Community Organisation
 Leverage underpinning work on similar
problems in Computer Science: Object
technology, software engineering,
distributed systems
 Link with related areas in Computer
Science dealing with different problems:
Artificial life, uncertainty in AI,
mathematical modelling
Community Organisation
 Extend and deepen links with other
disciplines: Economics, logic,
philosophy, sociology, etc
 Encourage industry take-up: Prototypes,
early adopters, case-studies, best
practice, early training
Existing software technology
 Build bridges with distributed systems,
software engineering and object technology.
 Develop agent tools and technologies on
existing standards.
 Engage in related (lower level)
standardisation activities (UDDI, WSDL,
 Clarify relationships between agent theories
and abstract theories of distributed
Different problems from related
 Build bridges to artificial life, robotics,
Uncertainty in AI, logic programming
and traditional mathematical modelling.
 Develop agent-based systems using
hybrid approaches.
 Develop metrics to assess relative
strengths and weakness of different
Prior results from other
 Maintain and deepen links with
economics, game theory, logic,
philosophy and biology.
 Build new connections with sociology,
anthropology, organisation design,
political science, marketing theory and
decision theory.
Encourage agent deployment
 Build prototypes spanning organisational
boundaries (potentially conflicting).
 Encourage early adopters of agent
technology, especially ones with some risk.
 Develop catalogue of early adopter case
studies, both successful and unsuccessful.
 Provide analyses of reasons for success and
failure cases.
Encourage agent deployment
 Identify best practice for agent oriented
development and deployment.
 Support standardisation efforts.
 Support early industry training efforts.
 Provide migration paths to allow smooth
evolution of agent-based solutions,
from today’s solutions,
Application Opportunities
Application Opportunities
 Ambient Intelligence
 Bioinformatics and Computational
 Grid Computing
 Electronic Business
 Simulation
 Semantic Web
Ambient Intelligence
 Pillar of European Commission’s IST vision
 Also developed by Philips in long-term vision
 Three parts
• Ubiquitous computing
• Ubiquitous communication
• Intelligent user interfaces
 Thousands on mobile and embedded devices
interacting to support user-centred goals and
Ambient Intelligence
 Suggests a component-oriented world
populated by agents
 Demands
• Virtual organisations
• Infrastructure
• Scalability
 Information explosion in genomics and
 Distributed resources include databases
and analysis tools
 Demands automated information
gathering and inference tools
 Open, dynamic and heterogeneous
 Examples: Geneweaver, myGrid
Grid Computing
 Support for large scale scientific endeavour
 More general applications with large scale
information handling, knowledge
management, service provision
 Suggests virtual organisations and agents
 Future model for service-oriented
Electronic Business
 Agents currently used in first stage –
merchant discovery and brokering
 Next step is real trading – negotiating deals
and making purchases
 Potential impact on the supply chain
 Rise in agent-mediated auctions expected
• Agents recommend
• But agents do not yet authorise agreements
Electronic Business
 Short term: travel agents, etc
• TAC is a driver
 Long term: full supply chain integration
 At start of 2001, there were
• 1000 public eMarkets
• 30,000 private exchange
 Education and training
 Scenario exploration
 Entertainment
The Two Towers
 Thousands of agents simulated using
the MASSIVE system
 Realistic behaviour for battle scenes
 Initial versions included characters
running away!
 Previous use of computational
characters did not use agent
behaviour (eg Titanic).
Current State
 Pivotal role in contributing to broader visions
of Ambient Intelligence, Grid Computing,
Semantic Web, etc.
 European strength is broad and deep
 Still requires integration, needs to avoid
fragmentation, needs effective coordination
 Needs to support industry take-up and
For more information ...
Dr Michael Luck
Department of Electronics and
Computer Science
University of Southampton
Southampton SO17 1BJ
United Kingdom
Feedback sought: please send feedback!
The Book
The CD
The Agent Portal